Breast Tumor Diagnosis using Machine Learning with Microwave Probes

In this paper, we propose a detection technique that combines a machine learning modality with microwave near-field probes for breast tumor diagnosis. The proposed technique uses a highly sensitive microwave probe to identify differences between normal and abnormal breasts. Distinguishing between healthy and non-healthy breast based on estimating the differences in the reflection coefficient of the probe response for both normal and abnormal cases. Machine learning techniques are applied to accentuate the variance in the sensor's responses for both healthy and tumorous cases. We investigated the detection of breast tumors if a woman has different breast sizes and she has an abnormality in one of them. We show that for two different breast phantom sizes, one with a tumor and one without, the sensor provides reliable detection. Simulation results of ninety different-size realistic breast phantoms (45 healthy breasts and 45 tumorous breasts) show that the proposed system provides highly encouraging reliable detection probability.